AI data centers have moved from a technical infrastructure story to a household cost story. As tech companies describe plans for large new facilities, many consumers are looking at the electric grid and asking a simple question: who pays when demand rises?
A report commissioned by solar installer Sunrun found that 80% of consumers are worried about the effect of data centers on their utility bills. That concern is not emerging in a vacuum. The source article points to a grid where demand patterns have changed, supply options face constraints, and AI has become an especially visible target.
Why utility bills are now part of the AI debate
For more than a decade, electricity demand in the United States was steady, according to the U.S. Energy Information Administration (EIA). That stability has shifted over the last five years as commercial users, including data centers, and industrial users began using more power.
The numbers show why the issue has become harder to ignore. Commercial users recorded annual growth of 2.6%, while industrial users rose 2.1%. Residential use grew much more slowly, at 0.7% annually.
That does not mean AI data centers are the only force behind higher demand. Industrial users have also been drawing more from the grid. But data centers are highly visible, and the AI boom has made them a symbol of the strain consumers fear could eventually show up on monthly bills.
Data centers are taking a bigger share of U.S. power
Data centers today consume about 4% of the electricity generated in the United States. That is more than double their share in 2018, according to the source article.
The expected growth is larger still. Lawrence Berkeley National Laboratory forecasts that data center consumption will rise to 6.7% to 12% by 2028. Even at the low end, that would represent a meaningful increase in the share of national electricity used by these facilities.
The core tension is straightforward: AI data centers need large amounts of electricity, and the pace of new demand is putting pressure on the systems that generate and deliver power. When consumers hear about massive new facilities, they are connecting those announcements to a basic household concern: higher utility bills.
The issue also intersects with public opinion about AI itself. A Pew survey cited in the source article found that more people are concerned about the technology than excited about it. The source article also notes that many employers have used AI as a way to cut headcount rather than improve or augment employee productivity. That context matters because rising energy prices could turn existing unease into a broader backlash.
Renewables have helped, but their path is uncertain
So far, electricity generation has been able to meet demand with help from new solar, wind, and grid-scale battery storage. Big tech companies have been signing large deals for new utility-scale solar, especially because solar offers low cost, modularity, and speed to power.
Solar also has a practical timing advantage for data center developers. Solar farms can begin delivering electricity before they are fully completed, and a new project typically takes around 18 months to complete.
The EIA expects renewables to dominate new generating capacity through at least the next year. But the source article notes that experts expect a Republican repeal of key parts of the Inflation Reduction Act to hamper renewable growth beyond 2026.
That creates a difficult planning problem. If demand from AI data centers and other large users continues to rise, but the fastest-growing sources of new power slow down, the grid could face tighter conditions. Consumers do not need to follow every detail of energy policy to understand the risk: if supply additions struggle to keep pace with demand, bills become a natural point of concern.
Natural gas is not filling the gap quickly
Natural gas is another energy source favored by data center operators, but it has not solved the timing problem. Production has been rising, yet most new supplies have gone toward feeding exports rather than the domestic market.
The contrast is sharp. Consumption by electricity generators rose by 20% between 2019 and 2024, while exporters consumed 140% more. That means higher production has not translated cleanly into the domestic power support that data center developers might want.
New natural gas power plants also take time. According to the International Energy Agency, they take around four years to complete. A backlog of turbines used by gas-fired power plants has made the situation more difficult, with manufacturers quoting delivery dates up to seven years out.
The source article adds that newly announced production capacity is unlikely to change things. For data center developers, that leaves few quick answers. Renewables may face policy pressure, while natural gas plants and equipment are slow to arrive.
Why AI may become the public face of higher energy costs
The energy story is broader than AI. Industrial users are also drawing more electricity, and the grid is responding to multiple demand sources at once. Still, AI data centers have become the headline example because they are expanding quickly and are closely tied to the technology industry’s current priorities.
That visibility matters. If consumers already feel cautious about AI, and if they also worry that data centers may raise utility bills, the technology becomes associated not only with workplace disruption but also with household costs.
The result is a political and consumer risk for the companies building AI infrastructure. Data centers are essential to the services tech companies want to offer, but their power needs are now part of a public debate over electricity demand, energy prices, and who benefits from the next wave of computing.
The facts in the source article point to a clear conclusion: the AI buildout is no longer only about chips, models, and software. It is also about the grid. Until developers can show that new power demand will be met without squeezing consumers, utility bills will remain central to the argument over AI data centers.